What does my group do?

Collaborative and exploratory analysis

• Data transformation and modeling: data smoothing, region finding (R/Bioconductor: Bsmooth, minfi)
• Exploration: search by gene, search by overlap
• Contextual analysis: overlap with other data (our own, other labs, UCSC, ensembl)

• We have unprecedented ability to measure
• and lots of publicly available data to contextualize it

Integrative, visual and computational exploratory analysis of genomic data

• Browser-based
• Interactive
• Integration of data
• Reproducible dissemination
• Communication with R/Bioconductor: epivizr package

Creativity in exploration

We are building software applications to support creative exploratory analysis of large genome-wide datasets…

Summarization: summarize integrated measurements (computed on data subsets)

Statistically-guided exploration: Calculate a statistic of interest

# Get tumor methylation base-pair data
m <- assay(se)[,"tumor"]

# Compute regions with highest variability across cpgs
region_stat <- calcWindowStat(m, step=25, window=80, stat=rowSds)
s <- region_stat[,"stat"]

Explore data based on statistic

What's around the regions with highest across CpG variability?

# get locations in decreasing order
o <- order(s, decreasing=TRUE)
indices <- region_stat[o, "indices"]
slideShowRegions <- rowRanges(se)[indices] + 1250000L
mgr$slideshow(slideShowRegions)

dynamically extensible: Easily integrate new data types and add new visualizations.

• Based on classic "three-table" design in genomic data analysis
• Data providers define coordinate space

Visualization design goals

• Context
• Integrate and align multiple data sources; navigate; search
• Connect: brushing
• Encode: map visualization properties to data on the fly
• Reconfigure: multiple views of the same data

Visualization goals

• Data
• Select and filter: tight-knit integration with R/Bioconductor;
• (current work) filters on visualization propagate to data environment
• Model
• New 'measurements' the result of modeling; perhaps suggested by data context

One interpretation of Big Data is Many relevant sources of contextual data

• Easily access/integrate contextual data
• Driven by exploratory analysis of immediate data
• Iterative process
• Visual and computational exploration go hand in hand

What is the measurement?

What is the measurement?

Features:

• Analysis units (features) unknown a priori
• High levels of sparsity
• Standard normalization methods don't work well
• Confounders, i.e. study site, countries, etc.
• Large number of features (Type 1 error control)

• Diarrheal study consisting ~1000 samples (now ~3000).
• Interesting microbiome for four countries / through ages
• Novel normalization and differential abundance testing framework for marker-gene surveys

Hierachically organized features

Hierarchically organized features

Defining the measurement unit of analysis

Not just features, but samples may be hierarchically organized

• Underlying idea: cut in the tree defines unit of analysis

• Visual design challenges: how to support effective exploration of cuts (persistence, consistency, density)

• Data challenges: how to support efficient exploration of cuts (graph database backend for contextual data)

• Engineering challenges: when do analysis patterns become interaction modes

• Systems for interactive (and creative) data exploration and analysis of epigenomic and metagenomic data
• Design exploration of hierarchical domains with statistical analysis as the ultimate goal
• Collaborative, reproducible, with close connection with R/Bioconductor (metagenomeSeq, metagenomeFeatures, metavizr)

Acknowledgements

Justin Wagner, Jayaram Kancherla (CBCB)
Florin Chelaru (now at Twinfog), Joseph Paulson (now at Harvard)
Mihai Pop (CBCB) Feinberg Lab & K. Hansen (JHU), R. Irizarry (Harvard) HMP2 Project (Xavier and Huttenhower, Harvard)

Funding: NIH, Genentech, Gates Foundation

More information

http://hcbravo.org
@hcorrada